Spin Equilibria in Monomeric Manganocenes ... - ACS Publications

Organometallics 2009, 28, 2005–2019

2005

Articles Spin Equilibria in Monomeric Manganocenes: Solid-State Magnetic and EXAFS Studies Marc D. Walter, Chadwick D. Sofield, Corwin H. Booth, and Richard A. Andersen* Department of Chemistry and Chemical Sciences DiVision of Lawrence Berkeley National Laboratory, UniVersity of California, Berkeley, California 94720 ReceiVed September 22, 2008

Magnetic susceptibility measurements and X-ray data confirm that tert-butyl-substituted manganocenes [(Me3C)nC5H5-n]2Mn (n ) 1, 2) follow the trend previously observed with the methylated manganocenes; that is, electron-donating groups attached to the Cp ring stabilize the low-spin (LS) electronic ground state relative to Cp2Mn and exhibit higher spin-crossover (SCO) temperatures. However, introducing three CMe3 groups on each ring gives a temperature-invariant high-spin (HS) state manganocene. The origin of the high-spin state in [1,2,4-(Me3C)3C5H2]2Mn is due to the significant bulk of the [1,2,4(Me3C)3C5H2]- ligand, which is sufficient to generate severe inter-ring steric strain that prevents the realization of the low-spin state. Interestingly, the spin transition in [1,3-(Me3C)2C5H3]2Mn is accompanied by a phase transition resulting in a significant irreversible hysteresis (∆Tc ) 16 K). This structural transition was also observed by extended X-ray absorption fine-structure (EXAFS) measurements. Magnetic susceptibility studies and X-ray diffraction data on SiMe3-substituted manganocenes [(Me3Si)nC5H5-n]2Mn (n ) 1, 2, 3) show high-spin configurations in these cases. Although tetra- and hexasubstituted manganocenes are high-spin at all accessible temperatures, the disubstituted manganocenes exhibit a small low-spin admixture at low temperature. In this respect it behaves similarily to [(Me3C)(Me3Si)C5H3]2Mn, which has a constant low-spin admixture up to 90 K and then gradually converts to high-spin. Thermal spin-trapping can be observed for [(Me3C)(Me3Si)C5H3]2Mn on rapid cooling. Introduction The molecular and electronic structure of manganocenes have been of interest since (C5H5)2Mn, Cp2Mn, was first prepared by Wilkinson1,2 and Fischer3 more than 50 years ago. In the solid state, (C5H5)2Mn is a chain polymer quite unlike the typical sandwich structure of D5h and D5d symmetry,4 although in the gas phase Cp2Mn has a sandwich structure like Cp2Fe.5,6 As originally reported, the magnetic behavior of Cp2Mn was not readily understood, until its polymeric structure was determined. A plot of the magnetic moment as a function of T shows that the individual Mn(II) centers are antiferromagnetically coupled in the temperature range 90-432 K above which it undergoes an abrupt phase transition to a pink form, where its magnetic moment is consistent with an isolated S ) 5/2 paramagnet, i.e., high-spin (HS) Mn(II). When diluted in Cp2Mg, Cp2Mn behaves as an isolated S ) 5/2 paramagnet from 90 to 500 K. The substituted derivative (MeC5H4)2Mn shows similar behavior in its magnetic susceptibility.7 The gas-phase electron diffraction data of (MeC5H4)2Mn have been interpreted as consisting of two monomeric spin isomers, where the ratio of high-spin * Corresponding author. E-mail: [email protected]. (1) Wilkinson, G.; Cotton, F. A. Chem. Ind. (London) 1954, 11, 307– 308. (2) Wilkinson, G.; Cotton, F. A. Prog. Inorg. Chem. 1959, 1, 1. (3) Fischer, E. O.; Jira, R. Z. Naturforsch. 1954, 9B, 618–619. (4) Grebenik, P.; Grinter, R.; Perutz, R. N. Chem. Soc. ReV. 1988, 17, 453–490. (5) Haaland, A. Top. Curr. Chem. 1975, 53, 1–23. (6) Haaland, A. Inorg. Nucl. Chem. Lett. 1979, 15, 267–269. (7) Wilkinson, G.; Cotton, F. A.; Birmingham, J. M. J. Inorg. Nucl. Chem. 1956, 2, 95–113.

(Mn-C(av) ) 2.43 Å) to low-spin (Mn-C(av) ) 2.14 Å) is 60:40 at 100 °C.8,9 This gas-phase structural study supports the deductions of earlier gas-phase photoelectron spectroscopy (PES) studies that (MeC5H4)2Mn is a spin equilibrium molecule in the gas phase, while Cp2Mn exists predominantly as the highspin isomer.10 The EPR studies of Ammeter and Maki have shown that Cp2Mn and (MeC5H4)2Mn in frozen solution and doped in a host matrix of Cp2M or (MeC5H4)2M (M ) Mg, Fe), respectively, behave as spin-crossover molecules, where a low-spin (LS) state is stabilized by about 0.5 kcal mol-1.11-14 The low-spin isomer (2E2g in D5h symmetry) has a doubly degenerate ground state and behaves as a dynamic Jahn-Teller molecule. The nature of the potential energy surface of the dynamic Jahn-Teller distortion depends upon the substituents on the Cp ring and the host metallocene, which has led to the study of various ring substituted compounds. The only other manganocene whose electronic structure has been studied extensively is (Me5C5)2Mn.15-17 The magnetic moment in the (8) Almenningen, A.; Haaland, A.; Samdal, S. J. Organomet. Chem. 1978, 149, 219–229. (9) Evans, S.; Green, M. L. H.; Jewitt, B.; King, G. H.; Orchard, A. F. J. Chem. Soc., Faraday 2 1974, 70, 356–376. (10) (a) Green, J. C.; Decleva, P. Coord. Chem. ReV. 2005, 249, 209. (b) Green, J. C. Faraday Discuss 2003, 124, 453–455. (c) Green, J. C.; Burney, C. Polyhedron 2004, 23, 2915–2919. (11) Ammeter, J. H.; Bucher, R.; Oswald, N. J. Am. Chem. Soc. 1974, 96, 7833–7835. (12) Ammeter, J. H. J. Magn. Reson. 1978, 30, 299–325. (13) Ammeter, J. H.; Zoller, L.; Bachmann, J.; Baltzer, P.; Gamp, E.; Bucher, R.; Deiss, E. HelV. Chim. Acta 1981, 64, 1063–1082. (14) Switzer, M. E.; Wang, R.; Rettig, M. F.; Maki, A. H. J. Am. Chem. Soc. 1974, 96, 7669–7674.

10.1021/om800922j CCC: $40.75  2009 American Chemical Society Publication on Web 03/06/2009

2006 Organometallics, Vol. 28, No. 7, 2009

Walter et al.

Table 1. Characterization Data for Substituted Maganocenes compound (Cp′)

color

C5H5 MeC5H4 C5Me5 Me3CC5H4 Me3SiC5H4 1,3-(Me3C)2C5H3 1,3-(Me3Si)2C5H3 1,2,4-(Me3C)3C5H2 1,2,4-(Me3Si)3C5H2 1,3-(Me3C)(Me3Si)C5H3

Tsub (°C)a

mp (°C)

amber amber orange red yellow red ivory light yellow ivory orange

Mn-C (av) (Å)

H NMR (δ)c

1

d

172-173 62-64 292 59-60 2σ(Fo2) 91 0.937-0.680 0.0366 0.0959 0.0467 1.052 0.299/-0.517

C26H42Mn 409.54 Pccn 11.693(1) 12.317(1) 32.877(1) 90 4734.94(15) 8 1.149 0.56 0.22 × 0.18 × 0.12 138(2) ω, 23.25 17 641 3401, 0.1165 2382, Fo2 > 2σ(Fo2) 257 0.992-0.774 0.0687 0.1287 0.1068 1.177 0.263/-0.477

C22H42MnSi4 473.86 P21/c 10.727(1) 12.971(1) 20.481(1) 97.316(1) 2826.49(9) 4 1.114 0.64 0.25 × 0.20 × 0.05 137(2) ω, 25.57 12 447 4769, 0.0716 3360, Fo2 > 2σ(Fo2) 256 0.990-0.859 0.0722 0.1900 0.0956 1.043 1.544/-0.530

C34H58Mn 521.74 P21/n 19.326(1) 17.640(1) 20.377(1) 112.483(3) 6418.9(2) 8 1.080 0.42 0.28 × 0.23 × 0.12 175(2) ω, 24.79 27 459 10 450, 0.0705 6595, Fo2 > 2σ(Fo2) 667 0.950-0.889 0.0491 0.1068 0.0978 0.981 0.390/-0.612

C28H56MnSi6 618.22 P21/a 18.5652(18) 22.341(2) 19.5687(19) 108.735(2) 7686.5(13) 8 1.068 0.55 0.36 × 0.30 × 0.24 131(2) ω, 24.74 33 864 12 656, 0.0642 8443, Fo2 > 2σ(Fo2) 694 0.880-0.837 0.0532 0.1275 0.0934 1.007 0.624/-0.623

R1 ) ∑||Fo| - |Fc||/∑|Fo|. b wR2 ) [(∑w(||Fo|2 - |Fc||2)2/∑w|Fo|4)]1/2.

solid state (5-100 K) shows it to be low-spin, and its EPR spectrum in frozen glasses and solid solutions of (Me5C5)2Fe show that the ground state is 2E2g, consistent with the gas-phase PES data. The room-temperature solid-state X-ray crystal structure of (Me5C5)2Mn with the Mn-C(av) distances of 2.11 Å shows it is low-spin, but there are several distortions, which have been ascribed to a static Jahn-Teller distortion.16 Herein, we report the synthesis, molecular structure, and electronic properties of a series of trimethylsilyl- and tert-butyl-substituted manganocenes.

Results and Discussion Synthesis. The synthesis of the manganocenes described in this paper (Table 1) generally uses the corresponding magnesocenes and either MnBr2 or MnI2(thf)2 in tetrahydrofuran except for Me5C5, C5H5, MeC5H4, (Me3Si)C5H4, (Me3Si)3C5H2, and (Me3C)3C5H2. In these examples the transfer reagent is either the corresponding potassium or sodium salts of the substituted (15) Cauletti, C.; Green, J. C.; Kelley, M. R.; Powell, P.; Tilborg, J. V.; Robbins, J.; Smart, J. J. Electron Spectrosc. Relat. Phenom. 1980, 19, 327– 353. (16) Freyberg, D. P.; Robbins, J. L.; Raymond, K. N.; Smart, J. C. J. Am. Chem. Soc. 1979, 101, 892–897. (17) Smart, J. C.; Robbins, J. L. J. Am. Chem. Soc. 1978, 100, 3936– 3937.

cyclopentadiene, since the magnesocene reagent did not react cleanly under a variety of conditions investigated with the aforementioned cyclopentadienyl derivatives. The manganocenes are soluble in pentane, and they are readily purified by crystallization from that solvent. They all sublime from 50 to 80 °C depending on the substituent under dynamic diffusion pump vacuum, except (Me3Si)C5H4, which is a liquid that distills at 92-93 °C in diffusion pump vacuum, and they all give molecular ions in their EI mass spectra. The melting behavior is rather interesting since the melting points increase in the order (Me3E)C5H4 < (Me3E)2C5H3 , (Me3E)3C5H2, where E ) Si or C. Curiously the melting point of [(Me3C)(Me3Si)C5H3]2Mn is close to the arithmetic mean of [(Me3C)2C5H3]2Mn and [(Me3Si)2C5H3]2Mn. The colors of the manganocenes are yellow in the case of (Me3Si)C5H4, (Me3Si)2C5H3, (Me3Si)3C5H2, and (Me3C)3C5H2, while the others are orange or red. Solid-State Crystal Structures. The crystal structures of [(Me3Si)2C5H3]2Mn, [(Me3Si)3C5H2]2Mn, [(Me3C)C5H4]2Mn, [(Me3C)2C5H3]2Mn, and [(Me3C)3C5H2]2Mn are determined at low temperature (see Table 2 for details), while that of [(Me3Si)C5H4]2Mn was reported some years ago.19 Thus, the solid-state structures of the manganocenes described in this paper (18) Bu¨nder, W.; Weiss, E. Z. Naturforsch. 1978, 33b, 1237–1238. (19) Hebendanz, N.; Ko¨hler, F. H.; Mu¨ller, G.; Riede, J. J. Am. Chem. Soc. 1986, 108, 3281–3289.

Spin Equilibria in Monomeric Manganocenes

Organometallics, Vol. 28, No. 7, 2009 2007

are known, except for the mixed-ring metallocene [(Me3C)(Me3Si)C5H3]2Mn; several attempts to grow suitable crystals for an X-ray diffraction study by either slow sublimation or crystallization failed. The ORTEPs are given in Figures 1-5, and the bond distances and angles are given in the figure captions (see Supporting Information for details). [(Me3C)C5H4]2Mn has C2h symmetry, the rings are staggered, and the CMe3 groups point away from each other to reduce intramolecular interactions. The compound is isostructural and isomorphous to its magnesium analogue,20 whose ionic radius is between that of high-spin and low-spin Mn2+.21 A similar geometry was reported by Ko¨hler for the SiMe3 analogue.19 [(Me3C)2C5H3]2Mn and [(Me3Si)2C5H3]2Mn have idealized C2 symmetry, and the Cp rings are essentially eclipsed and coplanar. In [(Me3C)3C5H2]2Mn and [(Me3Si)3C5H2]2Mn the Cp rings are essentially eclipsed, the rings are coplanar, and the molecule Figure 3. ORTEP of [1,3-(Me3Si)2C5H3]2Mn (50% probability ellipsoids). Selected distances (Å) and angles (deg): Mn-C (range) 2.322(4)-2.408(4), Mn-Cpcentroid 2.04, Cpcentroid-Mn-Cpcontroid ) 166. Average displacement of quaternary carbon from ring plane is 0.06 Å.

Figure 1. ORTEP of [(Me3C)C5H4]2Mn (50% probability ellipsoids). Selected distances (Å) and angles (deg): Mn-C (range) 2.117(2)-2.187(2), Mn-Cpcentroid 1.77, cyclopentadienyl rings are related by inversion, Cpcentroid-Mn-Cpcontroid ) 180. Average displacement of quaternary carbon from ring plane is 0.08 Å.

Figure 2. ORTEP of [1,3-(Me3C)2C5H3]2Mn (50% probability ellipsoids). Only one of the unique molecules is shown. Selected distances (Å) and angles (deg): Mn-C (range) 2.072(5)2.183(5), Mn-Cpcentroid 1.75, cyclopentadienyl rings are related by a C2 axis, Cpcentroid-Mn-Cpcontroid ) 175. Average displacement of quaternary carbon from ring plane is 0.16 Å.

Figure 4. ORTEP of [1,2,4-(Me3C)3C5H2]2Mn (50% probability ellipsoids). Only one of the unique molecules is shown. Selected distances (Å) and angles (deg): Mn-C (range) 2.359(3)2.487(3), Mn-Cpcentroid 2.11, Cpcentroid-Mn-Cpcontroid ) 169. Average displacement of quaternary carbon from ring plane is 0.21 Å.

has idealized C2 symmetry. A notable feature is that the CMe3 and SiMe3 groups deviate from the plane defined by five ring carbons by 0.21 and 0.27 Å, respectively. Although the manganocenes with the same number of Me3C and Me3Si groups attached to the C5 ring have the same idealized symmetry and orientation of the ring substituents, the Mn-C and Mn-Cp(cent) distances differ greatly (Table 1) between these manganocenes. Thus, the Mn-C distances in (Me3C)C5H4 and (Me3C)2C5H3 are, on average, 0.24 Å shorter than those in the (Me3Si)C5H4 and (Me3Si)2C5H3 manganocenes. In marked contrast, the average Mn-C distances in (Me3C)3C5H2 of 2.43(4) Å are the same as those in (Me3Si)3C5H2 of 2.40(3) Å within 3σ. The differences in the bond distances show that these (20) Gardiner, M. G.; Raston, C. L.; Kennard, C. H. L. Organometallics 1991, 10, 3680–3686. (21) For the coordination number 6 the ionic radii are Mn2+ (HS) ) 0.83 Å, Mn2+ (LS) ) 0.67 Å, and Mg2+ ) 0.72 Å (Shannon, R. D. Acta Crystallogr. A 1976, 32, 751-767).


Figure 5. ORTEP of [1,2,4-(Me3Si)3C5H2]2Mn (50% probability ellipsoids). Only one of the unique molecules is shown. Selected distances (Å) and angles (deg): Mn-C (range) 2.350(4)2.470(4), Mn-Cpcentroid 2.07, Cpcentroid-Mn-Cpcontroid ) 165. Average displacement of quaternary carbon from ring plane is 0.27 Å.

manganocenes have different spin states, with the shorter distances associated with the low-spin isomer. These distances compare well to those obtained for the high-spin isomer in (MeC5H4)2Mn of 2.433(8) Å and the low-spin isomer of 2.144(12) Å in the gas electron diffraction study. The value of the Mn-C distance in (Me5C5)2Mn of 2.111(3) Å clearly indicates it is low-spin, which is supported by variabletemperature magnetic susceptibility data (see below). Thus, Mn-C(av) distances and the color (yellow to ivory or red) imply high-spin and low-spin, respectively (at the temperature of the observation). Solid-State Magnetic Susceptibility Studies. The spin-state deductions derived from the Mn-C distances are supported by the variable-temperature magnetic susceptibility studies on these manganocenes over a temperature range from 5 to 300 K. A high-spin manganocene has a 6A1g ground state (in D5d symmetry) and will have an electronic configuration (e2g2 a1g1 (e1g*)2) and the expected spin-only value of µeff is 5.92 µB.22 A low-spin manganocene has a 2E2g ground state with electronic configuration e2g3a1g2 and the expected effective magnetic moment is about 2.2 µB, since there will be an orbital contribution to the moment.23 The reported effective moment for [(Me2CH)4C5H]2Mn of 5.72 µB is temperature independent from 5 to 350 K, consistent with the high-spin state (S ) 5/2) deduced by the Mn-C bond lengths.24 The effective magnetic moment of (C5Me5)2Mn of 2.17 µB from 5 to 100 K17,25 is also consistent with lowspin S ) 1/2, as deduced from the Mn-C distances; the magnetic data over a wider temperature range, 5-560 K, is temperature independent; see Supporting Information for details (Figure S6). Thus, the magnetic moment as a function of temperature indicates the spin state for the manganocenes. The magnetic susceptibility data for the three Me3Sisubstituted manganocenes are plotted as µeff vs T, as shown in Figure 6. For [(Me3Si)2C5H3]2Mn and [(Me3Si)3C5H2]2Mn, the magnetic moment, µeff, is independent of temperature over the temperature range studied, and the value of µeff is ∼5.9 µB, consistent with a high-spin state throughout the temperature range. The plots of χ-1 vs T, available in the Supporting

Walter et al.

Figure 6. Solid-state magnetic moment vs T plots for [(Me3Si)nC5H5-n]2Mn (n ) 1, 2, 3).

Information, Figure S7, are linear in T, showing that these two manganocenes follow the Curie-Weiss law over the temperature range, χ ) C/(T - θ), and the θ constants are small, -1.6 and -1.2 K for [(Me3Si)2C5H3]2Mn and [(Me3Si)3C5H2]2Mn, respectively. The small value of θ implies that they are isolated paramagnets, and the negative value of θ indicates weak antiferromagetic spin interactions or zero-field splitting (ZFS). The behavior of [(Me3Si)C5H4]2Mn is the same as the other two compounds from 300 to 150 K, but below this temperature the value of µeff experiences a nonlinear decrease until 100 K, below which the effective magnetic moment is essentially temperature independent. This suggests that the population of the spin isomers changes in the temperature regime 100-150 K, but that the populations are fixed at temperatures below 100 K and above 150 K. The decrease in µeff below about 20 K in all samples can be attributed to both weak intermolecular antiferromagnetic coupling and ZFS of the S ) 5/2 ground state.26 The magnetic behavior of the three Me3C-substituted derivatives is widely different and rather more interesting than [(Me3Si)2C5H3]2Mn and [(Me3Si)3C5H2]2Mn. Figure 7 shows plots of µeff as a function of temperature to 300 K for the three manganocenes. The behavior of [(Me3C)3C5H2]2Mn is like that found for the [(Me3Si)3C5H2]2Mn and [(Me3Si)2C5H3]2Mn derivatives; namely, µeff is independent of temperature from 5 to 300 K, and θ ) -1.9 K is similar to that found for [(Me3Si)3C5H2]2Mn and [(Me3Si)2C5H3]2Mn. The magnetic moment and the Mn-C bond distances are consistent with S ) (22) The effective magnetic moment is directly related to the spin and orbit quantum numbers as follows: µeff ) [4S(S + 1) + L(L + 1)]1/2. For ions with A and E ground-state terms, the magnetic properties can be assigned as predominantly spin in nature since most of the orbital angular moment has been quenched. Orbital quenching is nearly complete for ions with A1g ground terms, and as a result, the g values are very close to the free spin value of 2.0 and the zero-field splitting of the spin degeneracies is usually small. Partial quenching is apparent in the other ions with A and E ground-state terms, resulting in g value anisotropies, deviation of the moment from the spin-only value, and larger zero-field splittings of the spin multiplets. For manganocenes, the high-spin isomer with an 6A1g ground-state term exhibits a magnetic moment of µeff ) 2[2.5 × 3.5]1/2 ) 5.92 µB, since L ) 0. (23) Figgis, B. N. Introduction to Ligand Fields; J. Wiley & Sons: New York, 1966; Chapter 10. (24) Hays, M. L.; Burkey, D. J.; Overby, J. S.; Hanusa, T. P.; Sellers, S. P.; Yee, G. T.; Young, V. G. Organometallics 1998, 17, 5521–5527. (25) Robbins, J. L.; Edelstein, N. M.; Cooper, S. R.; Smart, J. C. J. Am. Chem. Soc. 1979, 101, 3853–3857.


Figure 7. Solid-state magnetic moment vs T plots for [(Me3C)nC5H5-n]2Mn (n ) 1, 2, 3).

5/2 in this example. The magnetic behavior becomes strongly temperature dependent, however, when the number of Me3C substituents is decreased by one, giving [(Me3C)2C5H3]2Mn, whose µeff vs T plot is shown in Figure 7. The χ-1 vs T and χT vs T plots are available in the Supporting Information, Figure S8. The value of µeff in the low-temperature range (5-200 K) is essentially independent of temperature and µeff is ∼2 µB, a value close to that found for (Me5C5)2Mn. In addition, since the Mn-C(av) bond distances for [(Me3C)2C5H3]2Mn and (Me5C5)2Mn are in the same range, these manganocenes have the same spin state, S ) 1/2. Inspection of Figure 7 shows that µeff is increasing slightly in the temperature regime from 5 to 250 K, at which point the slope appears to increase. When these results were originally obtained, it was not possible to obtain data at T > 300 K (due to the limitations of the sample container). However, using quartz tubes as sample containers (see Experimental Section for details) temperatures up to 700 K are now routinely accessible. The data from 5 to 400 K are shown in Figure 8, where it is clear that µeff changes quite rapidly over the temperature range 300 to 400 K, so that the value of µeff approaches that of 5.5 µB, i.e., the value expected for a S ) 5/2 molecule. Another feature in the µeff vs T plot is that cooling of the sample from 400 K does not trace the original heating curve until about 150 K; this manganocene shows a distinct hysteresis between initial heating and cooling curves of 16 K. Subsequent heating and cooling yields moments that follow the initial cooling curve with a smaller hysteresis (∆T ) 2 K), and annealing is finally observed as the original cooling curve is obtained after 5-10 cycles. Thus, [(Me3C)2C5H3]2Mn appears to be a spin-equilibrium molecule with a very wide transition temperature range and a hysteresis implying substantial reorganization of the individual molecules in the crystal as they change their spin state when approaching the melting point of 428-429 K. These results are most consistent with a manganocene whose ground state is 2E2g and the population of the 6A1g state gradually increasing up to 330 K. At this point [(Me3C)2C5H3]2Mn presumably undergoes a crystallographic phase transition with significant structural rearrangements. The phase transition is accompanied by a change in the electronic ground state to an essentially high-spin configuration, where the mole fraction of (26) Kahn, O. Molecular Magnetism; VCH: New York, 1993.


Figure 8. Solid-state magnetic moment vs T plot for [(Me3C)2C5H3]2Mn. These data were collected on initial heating and cooling.

the high-spin state is xHS ) 0.89, as indicated by a moment that approaches the spin-only value for manganese(II) at 400 K. Upon cooling, the population of the low-spin isomer increases until about 200 K when the population of the high-spin isomer is small. The hysteresis behavior is presumably due to the rate at which the spin isomers change population, which in turn is associated with the orientation of the individual molecules in the crystalline lattice, i.e., their space group symmetry. The initial heating and cooling cycle has a relatively large transition temperature associated with it, but becomes progressively smaller as the heating and cooling curves are cycled, presumably due to what is referred to as a decrease in lattice elasticity.27,28 The idea that the high- and low-spin isomers crystallize in different space groups is based upon a comparison between the crystal structures of [(Me3C)2C5H3]2Mn and [(Me3Si)2C5H3]2Mn, which are not isomorphous; [(Me3C)2C5H3]2Mn crystallizes in the orthorhombic space group, Pccn, and the molecule contains a crystallographically imposed C2-axis, whereas [(Me3Si)2 C5H3]2Mn crystallizes in the monoclinic space group, P21/c. Although no crystal structure data are available on the highspin isomer of [(Me3C)2C5H3]2Mn, it can be assumed that this isomer also adopts the more open but less ordered P21/c structure. The packing diagrams for [(Me3C)2C5H3]2Mn and [(Me3Si)2C5H3]2Mn are provided as Supporting Information, Figures S9 and S10. This hypothesis is supported by the close structural relation among [(Me3Si)2C5H3]2Mn, [(Me3Si)3 C5H2]2Mn, and [(Me3C)3C5H2]2Mn, which are isomorphous, isostructural, and high-spin manganocenes. Phase transitions that convert low-spin Pccn structures into high-spin P21/c structures are not without precedent and have been observed in coordination compounds.29 Furthermore the interconversion of the lowtemperature orthorhombic Cmca to the room-temperature monoclinic C2/c space group has been observed for (C5Me5)2Mn,30 although no spin-state change occurs. In the next section, we show that variable-temperature extended X-ray absorption fine structure (EXAFS) measurements are used to quantify the connection between magnetic moments and bond lengths, which support the qualitative inferences mentioned above. (27) Spiering, H.; Meissner, E.; Ko¨ppen, H.; Mu¨ller, E. W.; Gu¨tlich, P. Chem. Phys. 1982, 68, 65–71. (28) Adler, P.; Wiehl, L.; Meissner, E.; Ko¨hler, C. P.; Spiering, H.; Gu¨tlich, P. J. Phys. Chem. Solids 1987, 48, 517–525.


Walter et al.

Scheme 1

The magnetic behavior of [(Me3C)C5H4]2Mn is related to that of [(Me3C)2C5H3]2Mn except that the transition temperature is lower, about 200 K, and this manganocene is essentially a S ) 5/2 molecule at its melting point (332-333 K), where the color changes from red to yellow-orange on melting. Ammeter has shown by UV/vis spectroscopy that [(Me3C)C5H4]2Mn is a spinequilibrium molecule in solution.13 It is clear that the 1,3-R2C5H3 ligands induce vastly different magnetic properties on the resulting manganocene depending upon whether R2 is two Me3C or two Me3Si. It is then of interest to prepare a mixed manganocene and examine its magnetic behavior. The isomer [1-(Me3C)-3-(Me3Si)C5H3]2Mn was prepared, which exists as a mixture of diastereomers of idealized Cs (meso) or C2 (dl, rac) symmetry depending on the face to which the metal is bound (Scheme 1). The two optically active species (C2) are undistinguishable energetically and magnetically, but the meso-isomer has a different free energy and presumably a slightly different spin-crossover (SCO) profile. The plot of µeff vs T is shown in Figure 9 along with that of Me3Si, with which it is qualitatively similar. At high temperatures, µeff approaches the value expected for the high-spin isomer. The value of µeff gradually decreases on cooling to 100 K, then the value of µeff (4.21 µB) is essentially independent of temperature to 5 K. The temperature-independent region below 100 K depends upon the rate of cooling, as shown in Figures 10 and 11. When the sample is cooled rapidly (10 K/min) from 300 to 5 K (with or without field) and then warmed slowly, the moment is about 4.6 µB up to about 90 K, where it subsequently decreases to a minimum value at 105-110 K (µB ≈ 4.5 µB). The moment then smoothly increases to a maximum value of 5.9 µB at 300 K. However, if the sample is cooled slowly (1 K/min) from 300 to 5 K, the magnetic moment agrees well with the rapid-cooled data until about 105 K, where it decreases smoothly to µB ≈ 4.4 µB at 90 K, below which point it is independent of temperature. Although the last stages of the relaxation overlap with the thermal spin-crossover, it is possible to determine a thermal spin-crossover temperature, T1/2thermal, of 105 K as defined by Le´tard and co-workers31 (Figure 11, inset). It is, in fact, not unusual for magnetic susceptibility to be sensitive to the history of the sample. From coordination compounds it is known that a certain quantity of the high-spin population can get trapped on rapid cooling, because the lattice (29) Guionneau, P.; Le´tard, J.-F.; Yufit, D. S.; Chasseau, D.; Bravic, G.; Goeta, A. E.; Howard, J. A. K.; Kahn, O. J. Mater. Chem. 1999, 9, 985–994. (30) Augart, N.; Boese, R.; Schmid, G. Z. Anorg. Allgem. Chem. 1991, 595, 27–34. (31) Le´tard, J.-F.; Capes, L.; Chastanet, G.; Moliner, N.; Le´tard, S.; Real, J.-A.; Kahn, O. Chem. Phys. Lett. 1999, 313, 115–120.

Figure 9. Solid-state magnetic moment vs T plots for [(Me3C)(Me3Si)C5H3]2Mn and [(Me3Si)C5H4]2Mn.

Figure 10. Solid-state magnetic moment vs T plot for [1,3-(Me3C)(Me3Si)C5H3]2Mn.

does not have enough time to relax, a phenomenon known as “spin trapping”.32-39 Up to 6% of the high-spin population in [1,3-(Me3C)(Me3Si)C5H3]2Mn can be trapped on cooling 10 K/min, and the mole fractions xHS, trapped(5 K) ) 0.54 vs xHS, untrapped(5 K) ) 0.48 can be estimated from Figure 11. Relaxation curves in the temperature regime 80-102 K follow first-order kinetics, as expected from isolated metastable species, and are provided in the Supporting Information (Figure S11). Fitting the experi(32) Buchen, T.; Gu¨tlich, P.; Sugiyarto, K. H.; Goodwin, H. A. Chem.Eur. J. 1996, 2, 1134–1138. (33) Buchen, T.; Gu¨tlich, P.; Goodwin, H. A. Inorg. Chem. 1994, 33, 4573–4576. (34) Hauser, A. Chem. Phys. Lett. 1992, 192, 65–70. (35) Hauser, A. Comments Inorg. Chem. 1995, 17, 17–40. (36) Marchivie, M.; Guionneau, P.; Le´tard, J.-F.; Chasseau, D.; Howard, J. A. K. J. Phys. Chem. Solids 2004, 65, 17–23. (37) Roubeau, O.; deVos, M.; Stassen, A. F.; Burriel, R.; Haasnoot, J. G.; Reedijk, J. J. Phys. Chem. Solids 2003, 64, 1003–1013. (38) Ritter, G.; Ko¨nig, E.; Irler, W.; Goodwin, H. A. Inorg. Chem. 1978, 17, 224–228. (39) Yu, Z.; Liu, K.; Tao, J. Q.; Zhong, Z. J.; You, X. Z.; Siu, G. G. Appl. Phys. Lett. 1999, 74, 4029–4031.



Figure 11. χmT vs T plot for [1,3-(Me3C)(Me3Si)C5H3]2Mn. The inset shows the derivative of the curve after rapid cooling (10 K/min).

Figure 12. Magnitude of the Fourier transform (FT) of k3χ(k). Transform is from 2.5 to 10 Å-1, and Gaussian is narrowed by 0.3 Å-1.

mental curves to an expotential decay law gives satisfactory results for the rate constants k for the HS f LS conversion, which are gathered in Table 3. The ln k vs 1/T plot (Figure S12 in the Supporting Information) indicates that the HS f LS relaxation in that temperature regime behaves as a thermally activated process. The linear fit corresponding to a simple Arrhenius law (eq 1) is excellent (R2 ) 0.996), with A ) 20 ( 1 s-1, Ea ) 2.13 ( 0.02 kcal mol-1, and ∆Sq ) -50.4 cal mol-1 K-1 (see Supporting Information, Figure S12 for details). The value of 2.13 kcal mol-1 for the activation energy (Ea) is on the same order as found for [Fe(1-methyltetrazole)6](BF4), which also exhibits thermal spin trapping.40

Table 3. High-Spin (HS) f Low-Spin (LS) Relaxation Rates for [1,3-(Me3C)(Me3Si)C5H3]2Mn after Thermal Spin Trapping, Deduced from a Single-Expotential Law

( )

k ) A exp -

Ea kBT

(1)

EXAFS Study. The EXAFS technique can directly measure the local radial pair-distance distribution function around Mn atoms in these materials, thereby providing a structural parameter, such as distances, to compare with magnetic susceptibility data associated with the Mn spin state. The EXAFS technique has been widely successful in determining the local structure in Fe-based SCO species, for example, see refs 41-47. Unlike the studies of Fe compounds, special containment procedures are necessary that present a challenge for obtaining useful EXAFS data on the extremely air-sensitive manganocene compounds; see Experimental Section for details. In this study, the structural rearrangements in the SCOhysteresis sample [1,3-(Me3C)2C5H3]2Mn are investigated; data at two temperatures are shown in the Supporting Information, Figure S13. The data range is limited by a monochromator glitch at about 12.5 Å-1 that affects data at some temperatures more (40) Roubeau, O.; Stassen, A. F.; Gramage, I. F.; Codjovi, E.; Linare`s, J.; Varret, F.; Haasnoot, J. G.; Reedijk, J. Polyhedron 2001, 20, 1709– 1716. (41) (a) Gu¨tlich, P.; van Koningsbruggen, P. J.; Renz, F. Struct. Bonding (Berlin) 2004, 107, 24–75, and references therein. (b) Gu¨tlich, P.; Goodwin, H. A. Top. Curr. Chem. 2004 233, 1-47, and references therein. (42) Okamoto, K.; Nagai, K.; Miyawaki, J.; Kondoh, H.; Ohta, T. Chem. Phys. Lett. 2003, 371, 707–712. (43) Thuery, P.; Zarembowitch, J.; Michalowicz, A.; Kahn, O. Inorg. Chem. 1987, 26, 851–5. (44) Boca, M.; Vrbova, M.; Werner, R.; Haase, W. Chem. Phys. Lett. 2000, 328, 188–196.

T [K]

τ [s] × 103

k [s-1] × 10-5

80 85 90 95 100 102

30.03 ( 0.72 15.18 ( 0.22 7.82 ( 0.10 4.07 ( 0.07 1.93 ( 0.07 1.78 ( 0.04

3.33 ( 0.08 6.59 ( 0.09 12.79 ( 0.17 24.60 ( 0.42 51.91 ( 1.77 56.30 ( 1.17

than others; the samples were very difficult to align because they were strongly tapered. Therefore, the data are fit to 10 Å-1 when direct comparisons of temperature-dependent results are required. The limited data range does not affect the results, as shown below. The low-temperature r-space data show one dominant peak at 1.7 Å, followed by a smaller peak at 2.8 Å (Figure 12). The first peak is due primarily to 10 Mn-C paths (between 2.07 and 2.18 Å according to diffraction), although there is a small contribution by 6 Mn-H paths (∼2.8 Å). Somewhat surprisingly, it is necessary to include these Mn-H paths to allow for a reasonable S02 value (we use S02 ) 0.72 from CaMnO3).48 The peak at 2.8 Å is due to 4 Mn-C’s. The temperature dependence of the Fourier transforms (FTs) is very clear in the raw data, with a sharp drop in amplitude near 330 K. A comparison of the high-spin mole fractions, obtained from the relative amplitudes of the two near-neighbor Mn-C shells and from the solid-state magnetism, is excellent, Figure 13.49 Whereas the EXAFS study provides insight at the molecular level, the magnetic susceptibility reflects a bulk property; yet (45) Paulsen, H.; Grunsteudel, H.; Meyer-Klaucke, W.; Gerdan, M.; Grunsteudel, H. F.; Chumakov, A. I.; Ruffer, R.; Winkler, H.; Toftlund, H.; Trautwein, A. X. Eur. Phys J. B 2001, 23, 463–472. (46) Rudd, D. J.; Goldsmith, C. R.; Cole, A. P.; Stack, T. D. P.; Hodgson, K. O.; Hedman, B. Inorg. Chem. 2005, 44, 1221–1229. (47) Hannay, C.; Hubin-Franskin, M.-J.; Grandjean, F.; Briois, V.; Itie, J. P.; Polian, A.; Trofimenko, S.; Long, G. J. Inorg. Chem. 1997, 36, 5580– 5588. (48) Booth, C. H.; Bridges, F.; Kwei, G. H.; Lawrence, J. M.; Cornelius, A. L.; Neumeier, J. J. Phys. ReV. B 1998, 57, 10440–10454. (49) The X-ray absorption near-edge structure (XANES) region is consistent with the interpretation that the electronic state is best described as a mixture of the HS and LS configurations in the crossover region near 350 K; see Supporting Information, Figure S28, for further details.


Figure 13. Mole fraction of high-spin isomer vs T plot, obtained from the relative amplitudes of the two near-neighbor Mn-C shells by EXAFS and magnetic susceptibility studies.

Figure 14. ln K vs T-1 plot for high-spin, low-spin equilibrium for [1,3-(Me3C)2C5H3]2Mn determined by solid-state magnetic susceptibility after initial phase transformation (230-400 K). The full line represents a fit to a simple linear regression (see text). The obtained fit is moderately good (R2 ) 0.973) and gives the values of ∆H0 ) 3.8 ( 0.2 kcal/mol and ∆S0 ) 12.1 ( 0.5 cal/(mol K).

both methods confirm that a phase transition actually takes place and that it is the origin of the observed hysteresis behavior. After the transformation into the more open, probably “disordered” P21/c structure occurs, it is impossible to adopt the original structure on cooling; therefore the spin transition is no longer subject to constraints imposed by the higher symmetry structure and can proceed smoothly over a broader temperature regime. The substituted manganonocene [(Me2CH)3Me2C5)2Mn also exhibits an abrupt spin transition with a hysteresis at 167 K, but no further details are reported, and the molecular origin remains unknown.50 EPR Studies. The EPR spectra of manganocenes have been extensively studied, since the low-spin isomers have a 2E2g (50) Sitzmann, H.; Scha¨r, M.; Dormann, E.; Kelemen, M. Z. Allg. Anorg. Chem. 1997, 623, 1609–1613.

Walter et al.

Figure 15. Mole fraction high-spin vs T plot for high-spin, lowspin equilibrium for [1,3-(Me3C)2C5H3]2Mn determined by solidstate magnetic susceptibility after initial phase transformation (180-400 K). The full line represents a fit to a regular solution model (see text). ∆H0 ) 3.5 ( 0.2 kcal/mol; ∆S0 ) 10.6 ( 0.6 cal/(mol K).

ground state, which means they are Jahn-Teller molecules. The spectra are usually observed only at very low temperatures, since the spin-lattice relaxation times are long. For the high-spin metallocenes, the zero-field splitting parameter, D, is large (relative to the applied field) and a g-value of about 6 is obtained for MnF2, MnH2, MnO,25,51-53 and Cp2Mn in Cp2Mg.12 For the low-spin molecules, g| and g⊥ are observed at values of about 2 (Cp2Mn in Cp2Fe, (C5Me5)2Mn in methylcyclohexane glass and doped into (C5Me5)2Fe). In this study, the EPR spectra of a representative number of manganocenes are studied as frozen solutions at 4 K in a methylcyclohexane glass (see Supporting Information for details, Figures S14-S16). The manganocene [(Me3Si)2C5H3]2Mn is high-spin to 4 K, and it shows a featureless EPR spectrum with a geff value of 6.0. The lowspin manganocene [(Me3C)2C5H3]2Mn yields a spectrum with g| ) 2.57 and g⊥ ) 1.90. These two manganocenes follow the expected pattern, and the µeff values, determined from EPR spectra, are in agreement with those obtained from SQUID measurements. The EPR spectrum of the mixed manganocene [(Me3C)(Me3Si)C5H3]2Mn substantiates the contention that both spin isomers are present at T < 100 K, since a feature due to the S ) 5/2 state is observed at g ) 5.95 and features due to the S ) 1/2 state are observed, g| ) 2.88 and g⊥) 1.95. The EPR spectrum clearly supports the proposition that this manganocene exists in the two states, S ) 5/2 and S ) 1/2, and the quartet state is not present. Furthermore, the g values for the low-spin isomer can be used to calculate the magnetic moment,54 which agrees with the value measured, as shown in Table 4. Thermodynamics of SCO. Solid-State Studies. SolidState Magnetism. Deriving thermodynamic data from solidstate susceptibility measurements is rather complicated. In most (51) Van Zee, R. J.; Brown, C. M.; Weltner, W. Chem. Phys. Lett. 1979, 64, 325–328. (52) DeVore, T. C.; Van Zee, R. J.; Weltner, W. J. Chem. Phys. 1978, 68, 3522–3527. (53) Weltner, W., Magnetic Atoms and Molecules; Dover Publications: New York, 1983. (54) The magnetic moment of the low-spin isomer can be calculated based on EPR studies using the following equation: µLS ){1/3(g⊥2 + 2g|2)(S)(S + 1)}1/2.



Table 4. Magnetic Moments As Derived from Solid-State Magnetism and EPR Measurements compound

µHS (by SQUID)

µHS (by EPR)

µLS (by SQUID)

µLS (by EPR)

[1,3-(Me3C)2C5H3]2Mn [1,3-(Me3C)(Me3Si)C5H3]2Mn [1,3-(Me3Si)2C5H3]2Mn

n/a 5.89 ( 0.01 5.88 ( 0.01

n/a 5.86 ( 0.02 5.92 ( 0.01

1.91 ( 0.01 n/a n/a

1.86 ( 0.02 1.99 ( 0.02 n/a

Table 5. Thermodynamic Properties of the SCO in Various Manganocenes compound [(Me3C)C5H4]2Mn

[(Me3C)2C5H3]2Mn [(Me3C)(Me3Si)C5H3]2Mn a

method

K (300 K)

∆G 0(300 K) (kcal mol-1)

∆H0 (kcal mol-1)

∆S0 (cal mol-1 K-1)

χ (solid)a χ (solution)b UV-visb 1 H NMRc χ (solid)a

10.5 7.5 2.3

-1.4 -1.2 -0.5

0.7 0.7 3.2 17.3 47.4

0.2 0.3 -0.7 -1.7 -2.3

3.1 ( 0.2 2.0 3.7 2.6-3.2 3.8 ( 0.2 3.5 ( 0.2 2.7 ( 0.1 1.7 ( 0.1 3.1 ( 0.1

14.9 ( 0.4 10.5 13.9 9.6-12.0 12.1 ( 0.5 10.6 ( 0.6 11.2 ( 0.4 11.4 ( 0.5 18.0 ( 0.5

UV-visa χ (solid)a UV-visa

Τ1/2d [Κ] 211 194 266 314 327 241 148 174

This work. b Ref 13.13 c Ref 19. d T1/2 ) ∆H0/∆S0 (∆G0 ) 0)

cases the noninteracting-molecules approach is no longer valid, because the interaction between the different spin-carriers in the assembly has to be included; that is, the cooperativity has to be modeled. A variety of models have been put forward to account for these cooperativity effects, most notably the domain model and the regular solution model.26 An Arrhenius plot, i.e., the ln K vs T-1 function, provides information about the departure from ideal solution behavior; a linear function is characteristic of the absence of cooperative effects, and deviation from linearity increases with the extent of the interaction. As shown in Figure S17 in the Supporting Information, the Arrhenius plot for [1,3-(Me3C)2C5H3]2Mn deviates from linearity, indicating cooperative behavior. As shown in Figure 7, the magnetic moment, µeff, is more or less constant up to 210 K (1/T ) 0.0047) and then increases smoothly to 5.47 µB at 400 K (1/T ) 0.0025). This suggests that the population of the highspin state is small up to 210 K. To get a rough estimate of the thermodynamic constants for the equilibrium, only the temperature regime from 230 to 400 K is considered, since this is the temperature range in which the low-spin f high-spin actually occurs. This temperature range is shown in Figure 14. The solution model gives the plot shown in Figure 15.26 These models do not give fits of high accuracy; additional parameters to account for the interaction must be included, which makes the fit more flexible, but less meaningful in developing a molecular model for the intermolecular cooperativity. The regular solution model was applied to the spin-crossover in [(Me3C)C5H4]2Mn, whose thermodynamic properties were evaluated by Ammeter13 and Ko¨hler.19 The values for ∆H0 and ∆S0 for the solid state are in good agreement with the values obtained by solution studies (Table 5 and Figure S18 in the Supporting Information). Fitting the normalized mole fractions vs T for [1,3-(Me3C)(Me3Si)C5H4]2Mn to regular solution models gives a poor fit, although the resulting thermodynamic values are within the expected regime; the mole fraction vs T plot is available in the Supporting Information, Figure S19. Solution Studies. UV-Vis Studies. Ammeter evaluated the spin-equilibrium in [(Me3C)C5H4]2Mn by variable-temperature UV-vis spectroscopy.13 This technique also gives thermodynamic data for the SCO in the [1,3-(Me3C)2C5H3]2Mn and [1,3(Me3C)(Me3Si)C5H3]2Mn complexes. When the nearly colorless solution of [1,3-(Me3C)(Me3Si)C5H3]2Mn is cooled in liquid nitrogen, the orange color intensifies significantly, but on warming to room temperature, the color fades, implying a reversible equilibrium. The high-spin population is favored at

high temperature, whereas the low-spin contribution dominates at low temperature. In free energy terms, the entropy dominates at high temperatures and therefore favors the high-spin species. As the temperature is decreased, the T∆S0 term becomes smaller and the enthalpy term dominates the equilibrium. The facts that high-spin manganocenes are nearly colorless and low-spin species are orange to red allow the equilibrium to be studied using UV-vis spectroscopy. The low-spin complex [1,3-(Me3C)2C5H3]2Mn exhibits a strong d-d transition (2E2g f 2E1u)13 at λmax ) 430 nm. For the high-spin isomer, the transition is forbidden. The variable-temperature UV-vis experiment clearly shows an increase in the absorption at 430 nm as the sample is cooled to 170 K (Figure 16). Unfortunately at all temperatures accessible, the low-spin species is in equilibrium with the highspin species; so the concentration of high-spin and low-spin species at a given temperature is unknown, but the sum is equal to the total concentration. The solid-state magnetism indicates that the equilibrium between low- and high-spin [1,3(Me3C)2C5H3]2Mn shifts to high-spin above 250 K. The absorbance at 269 and 430 nm for high- and low-spin isomers, respectively, are temperature dependent, and the signal at 430

Figure 16. Variable-temperature UV-vis spectra for [1,3(Me3C)2C5H3]2Mn.


Walter et al.

Figure 18. Frontier orbital scheme for d5-metallocene with idealized D5d symmetry.

Figure 17. ln K vs T-1 plot for high-spin low-spin equilibrium of [1,3-(Me3C)2C5H3]2Mn determined by UV-vis spectroscopy. The thermodynamic values of ∆H0 ) 2.7 ( 0.1 kcal mol-1 and ∆S0 ) 11.2 ( 0.4 cal mol-1 K-1 are obtained by a least-squares fit (R2 ) 0.986).

nm shows a significant decrease in absorbance as the temperature is increased, whereas the feature at 230 nm increases with temperature. In order to extract the equilibrium constant, it is assumed that the extinction coefficient of [1,3-(Me3Si)2C5H3]2Mn is, to the first approximation, identical to the extinction coefficient for the high-spin species of [1,3-(Me3C)2C5H3]2Mn, since both species exhibit a sharp absorbance at 269 nm. This absorbance with an extinction coefficient is ) 25 L mol-1 cm-1 is temperature independent in [1,3-(Me3Si)2C5H3]2Mn, consistent with the solid-state magnetism, which is high-spin at all temperatures (see Supporting Information for details, Figure S20). This estimate is further corroborated by measuring the extinction of the 330 nm ( ) 2.12 L mol-1 cm-1) absorption, which is also temperature invariant in [1,3-(Me3Si)2C5H3]2Mn; the values obtained for the concentration of the high-spin form agree in both cases. The thermodynamic data of this spinequilibrium are determined by estimating the value of the extinction coefficient, at the lower temperature limit of the experiment. The absorbance of [1,3-(Me3C)2C5H3]2Mn at 430 nm changes little with temperature, and the populations of the isomers are determined from the absorbance data. Furthermore, in solution, the spin conversion is essentially the property of the individual, isolated molecule and the thermodynamics can be evaluated by assuming noninteracting molecules (eqs 2 and 3); the ln K vs T-1 plot is shown in Figure 17.

K)

xhs 0 ) e-(∆G /RT) 1 - xhs

ln K ) -

∆H0 ∆S0 + RT R

(2)

(3)

Evaluating the thermodynamic constants for [1,3-(Me3C)(Me3Si)C5H2]2Mn is more complicated, since the complex exists as a mixture of four different isomers that are divided into two pairs of enantiomers of C2 (rac) or Cs (meso) symmetry, respectively (see above). The two optically active species (C2) have the same free energy as do the meso-isomers, but the racand meso-isomers do not have identical free energies. This complication manifests itself in the fact that although the plot

Figure 19. Influence of substitution on the cyclopentadienyl ring on its frontier orbitals.56

of absorbance at 430 nm vs T of [1,3-(Me3C)(Me3Si)C5H2]2Mn shows that very little of the low-spin species is present at room temperature (see Supporting Information, Figure S21), the Arrhenius plot of ln K vs T-1 for [1,3-(Me3C)(Me3Si)C5H2]2Mn is significantly curved, suggesting that more than two species are involved in the equilibrium in contrast to [1,3(Me3C)2C5H3]2Mn (see Supporting Information, Figure S22). Since it is not possible to determine the population changes of the four isomers, a linear approximation of the curve, to a first approximation, gives enthalpy and entropy changes similar to those obtained previously, Table 5. Model that Rationalizes the Substituent Effects. Manganese(II) has a d5-electron configuration that gives rise to two possible d-electron configurations in D5d symmetry, viz., 2E2g (e2g3 a1g2) or 6A1g (e2g2 a1g1 (e1g*)2) low-spin or high-spin states, respectively. The choice depends upon the energy difference between the e2ga1g and the e1g* configurations (∆) and the spinpairing energy (P), Figure 18. In general, the results observed for the metallocenes described in this article are that Me3Si groups favor high-spin manganocenes and Me3C groups favor either low-spin or spinequilibrium manganocenes. This electronic effect can be traced to the classification of a Me3Si group as an electron-withdrawing substituent and the Me3C group as an electron-donating substituent deduced from EPR studies on the substituted cyclopentadienyl radicals. Figure 19 shows how these substituents change the relative energy of Ψs.55,56 In a monosubstituted cyclopentadienyl radical a substituent on C(1) does not change the energy of Ψa, because it lies on the nodal plane; however, Ψs is affected, since C(1) contains electron density. The pπ-electron density at each of the carbon atoms in the cyclopentadienyl ring is determined from the value (55) Kira, M.; Watanabe, M.; Sakurai, H. J. Am. Chem. Soc. 1980, 102, 5202–5207. (56) Barker, P. J.; Davies, A. G.; Tse, M. J. Chem. Soc., Perkin Trans. 2 1980, 941–948.


of the electron-nuclear coupling constant in the radical and is correlated with how the substituents change the relative energy of Ψa and Ψs. The EPR results show that π-electron-withdrawing groups stabilize Ψs, whereas π-electron-donating groups destabilize Ψs. In D5d, Ψs has e1g symmetry, which is the same symmetry as the dxz and dyz atomic orbitals used in metallocene bonding. Interaction of the e1g orbitals on the metal and the ligand fragments generates a bonding (e1g) and an antibonding (e1g*) combination. When R is electron donating, the energy of Ψs is closer in energy to that of the metal dxz and dyz orbitals, and the resulting molecular orbitals will be stabilized (e1g) and destabilized (e1g*) relative to when the substituent is H. The result is that ∆ is increased, resulting in a low-spin manganocene. Conversely, when R is electron withdrawing, ∆ decreases, resulting in a high-spin manganocene. This qualitative molecular orbital model is similar to that advanced by Ko¨hler.19 This model traces the control of spin states to electronic effects; however steric effects do play a role, particularly when the number of Me3C substituents is greater than 2; [1,2,4(Me3C)3C5H2]2Mn is high-spin at all the temperatures examined, a fact that is most reasonably ascribed to steric repulsion between the rings, which enforces longer Mn-C bond distances and lowers ∆. Steric effects are also suggested to rationalize the high-spin behavior of [(Me2CH)4C5H]2Mn.24,50

Conclusion The magnetic properties of manganocenes have interested chemists since the first manganocene, (C5H5)2Mn, was prepared and studied by Wilkinson1,7 and Fischer3 and their co-workers. Although (C5H5)2Mn is a zigzag polymer in the solid state, bulky substituents on the cyclopentadienyl rings result in monomers whose magnetic moments depend on their type and number. The two manganocenes described in detail in this article, [(Me3C)xC5H5-x]2Mn, x ) 1, 2, are classic textbook examples of spin-equilibrium molecules whose thermodynamic constants are determined in the solid state and in solution by magnetic susceptibility and UV-vis spectra as a function of temperature, respectively. However, these two physical techniques do not provide structural information, as the molecules change their spin states as the single-crystal X-ray structures are only done at one temperature. Molecular structural information is, however, provided by EXAFS data as a function of temperature on [(Me3C)2C5H3]2Mn, since the Mn-C bond distances and therefore the mole fraction of the high-spin isomer (or lowspin isomer) are determined directly. A plot of the mole fraction determined from the EXAFS data coincides with that obtained from the magnetic susceptibility data, as measured by χ, as a function of temperature, Figure 13. This correlation provides a direct measure of the molecular structural changes as the molecules in the solid-state ensemble change their spin state. The combination of all of these physical techniques studied in this article provides a detailed molecular level of understanding of [(Me3C)2C5H3]2Mn as it isomerizes.

Experimental Section General Comments. All reactions, product manipulations, and physical studies were carried out as previously described.57-59 (57) Lukens, W. W.; Matsunaga, P. T.; Andersen, R. A. Organometallics 1998, 17, 5240. (58) Walter, M. D.; Schultz, M.; Andersen, R. A. New J. Chem. 2006, 30, 238–246. (59) Walter, M. D.; Berg, D. J.; Andersen, R. A. Organometallics 2006, 25, 3228–3237.

Organometallics, Vol. 28, No. 7, 2009 2015 Magnetic measurements were conducted in a 7 T Quantum Design MPMS magnetometer utilizing a superconducting quantum interference device (SQUID). Between 10 and 25 mg of sample was sealed in evacuated quartz tubes held in place with ∼5 mg of quartz wool. This method provided a very small and reliable container correction, typically of about -2 × 10-5 emu/mol. The data were also corrected for the overall diamagnetism of the molecule using Pascal constants.60 For a more detailed description see ref 58. GC-MS analysis was performed on a Hewlett-Packard HP 6890 Series GC-System with a HP 5973 mass-selective detector. Routine UV-vis spectra were recorded on a Varian Cary 5G spectrometer. The following compounds were prepared as previously described: (Me3C)C5H5,61 [1,3-(Me3Si)2C5H3]2Mg,62[1,3-(Me3C)2C5H3]2Mg,63,64 NaNH2,65 and (Me3C)3C5H3.66 Dibutylmagnesium was purchased as a heptane solution from Aldrich, and its concentration was determined by titration. Variable-Temperature UV-Visible Studies. The variabletemperature UV-vis spectra were collected using a low-temperature apparatus of our own construction. The sample holder for an Ocean Optics ST 2000 fiber optic UV-vis spectrometer was connected to a copper coil arranged so that the coil could be immersed in a liquid N2 cold bath. During the experiment, dry N2 was allowed to flow through the copper coils and then through the cell holder. The sample holder was contained in an acrylic box that was purged with dry N2 to minimize water condensation. A quartz cuvette was fitted with a greaseless stopper that contained an opening, through which a thermocouple was inserted and sealed with epoxy. During a typical experiment, the cuvette containing the solvent (methylcyclohexane), with and without solute (for background spectra), was placed into the cell holder inside the acrylic box, and the container was purged with dry N2. The copper coil was placed inside a Dewar flask and cooled with liquid N2, while the coil was purged with dry N2. The temperature, which was monitored by the thermocouple in the cuvette, was regulated by the dry N2 flow rate. A typical experiment consisted of 20-50 data points in the temperature range of 30 to -110 °C over a time period of about 3 h. The collection time for each spectrum was 3-5 ms, and the spectra were collected at a given temperature until no change in absorbance was observed. EXAFS Studies. Preparing the [1,3-(Me3C)2C5H3]2Mn complex for X-ray absorption measurements required mixing about 17 mg of sample (corresponding to a change in absorption at the Mn K edge of about 1.5 absorption lengths) with dried boron nitride in an inert N2 atmosphere glovebox and loading it into a slotted, multiple-sample aluminum holder with lead-wire sealed aluminum windows. This holder guarantees the sample integrity during transportation in a container filled with nitrogen during transport to the Stanford Synchrotron Radiation Laboratory (SSRL) and handling in air (only about 1 min) while loading the sample holder into an evacuated LHe-flow cryostat. X-ray absorption data were collected on beam line 11-2 using an unfocused, uncollimated beam with energy resolution (98% pure (Me3C)2C5H4 (150 g, 0.841 mol, 70% yield). (1-Me3Si)(3-Me3C)C5H4. The magnesocene [(Me3C)C5H4]2Mg (74 g, 0.28 mol) was dissolved in tetrahydrofuran (100 mL), and freshly distilled trimethylsilyl chloride (71.1 mL, 60.84 g, 0.56 mol) was added dropwise. The reaction was noticeably exothermic, and a colorless precipitate formed before the addition was complete. The mixture was heated to reflux for 5 h to ensure complete conversion of the magnesocene. Filtration, removal of the solvent under reduced pressure, and distillation of the product at 38-41 °C in an oil pump vacuum yielded a colorless liquid (105 g, 0.54 mol, 96%). Anal. Calcd for C12H22Si: C, 74.17; H, 11.41. Found: C, 74.99; H, 11.96. The EI mass spectrum showed a parent ion at m/e ) 194 amu. 3-tert-Butyl(1-trimethylsilyl)cyclopentadiene has been prepared using another route.75 2,5,5-(Me3Si)3C5H4. The magnesocene [1,3-(Me3Si)2C5H3]2Mg (15 g, 33.8 mmol) was dissolved in tetrahydrofuran (100 mL), and freshly distilled trimethylsilyl chloride (8.6 mL, 7.34 g, 67.6 mmol) was added dropwise. The mixture was heated to reflux for 12 h. Cooling to room temperature, filtration, removal of the solvent under reduced pressure, and distillation of the product at 55-62 °C in an oil pump vacuum yielded a pale yellow liquid (16.6 g, 58.7 mmol, 87%). The product was analyzed by GC-MS and was typically >98% pure. The NMR spectroscopic data agree with those reported previously.76 Magnesocenes and Sodium Cyclopentadienides. [(Me3C)C5H4]2Mg. 1,1′-Di(tert-butyl)magnesocene, [(Me3C)C5H4]2Mg, was synthesized by treating mono(tert-butyl)cyclopentadiene (100 g, 0.82 mol) with a heptane solution of dibutylmagnesium (470 mL, 0.87 M, 0.41 mol). The solution became warm, and gas evolution was vigorous. The mixture was stirred at room temperature for 12 h. The solution was filtered, concentrated to a volume of 250 mL, and slowly cooled to -80 °C. Colorless crystals were isolated (69 g, 63%). A second crop of crystals (15 g) was obtained by concentrating the mother liquor and cooling to -20 °C. Overall yield: 84 g (0.315 mol, 77%). Mp: 93-95 °C. [(Me3C)C5H4]2Mg sublimes at 75-80 °C in an oil pump vacuum. The NMR spectroscopic data agree with those previously reported for this compound.20 (73) 232. (74) 2809. (75) (76)

Abel, E. W.; Moorhouse, S. J. Organomet. Chem. 1971, 29, 227– Venier, C. G.; Casserly, E. W. J. Am. Chem. Soc. 1990, 112, 2808– Okuda, J. Chem. Ber. 1989, 122, 1075–1078. Jutzi, P.; Sauer, R. J. Organomet. Chem. 1973, 50, C29-C30.

Spin Equilibria in Monomeric Manganocenes [(Me3Si)C5H4]2Mg. 1,1′-Bis(trimethylsilyl)magnesocene, [(Me3Si)C5H4]2Mg, was synthesized by treating mono(trimethylsilyl)cyclopentadiene (19.6 g, 0.142 mol) with a heptane solution of dibutylmagnesium (11.4 mL, 0.625 M, 0.071 mol). The mixture was heated at gentle reflux for 3 h. The solution was cooled to room temperature and filtered, and the solvent removed under reduced pressure, leaving behind a light yellow oil. In contrast to a previous report,62 pure [(Me3Si)C5H4]2Mg was obtained as a colorless, viscous liquid by distillation in a diffusion pump vacuum with a bath temperature of 80-100 °C (13.04 g, 0.04 mol, 57%). 1 H NMR (C6D6, RT, 400 MHz): δ 6.25-6.20 ppm (m, AA′BB′, H2,2′ and H3.3′, 4H, ring-CH), 0.26 ppm (s, 9H, Si(CH3)3). [(3-Me3C)(1-Me3Si)C5H3]2Mg. To 3-tert-butyl(1-trimethylsilyl)cyclopentadiene (110 g, 0.57 mol) was added a solution of dibutylmagnesium in heptane (320 mL, 0.87 M, 0.28 mol). The mixture was heated to reflux, and the butane evolution was monitored with an oil bubbler. After 12 h the reaction was complete, as determined by isolation of the system from the oil bubbler for 10 min without generation of pressure. The solution was filtered and slowly cooled to -80 °C. Colorless crystals were obtained (88 g). The mother liquor was concentrated to a volume of 50 mL and cooled to produce a second crop of crystals (6.4 g). Overall yield: 94.4 g, 0.23 mol, 82%. Mp: 107-108 °C (rev) [lit.: 99 °C77]. Sublimation: 95-100 °C (oil pump vacuum). IR (Nujol mull; CsI windows; cm-1): 1330w, 1290w, 1250s, 1205w, 1180m, 1170m, 1085m, 1055m, 1025w, 945m, 930m, 835s, 825s, 755s, 725w, 690m, 650w, 635m, 510m, 485m, 450w, 430m, 360w, 340w, 325w, 290w. 1H NMR (C6D6, 500 MHz): 0.301 (18H, s), 0.308 (18H, s), 1.30 (18H, s), 1.31 (18H, s), 6.16 (2H, dd, 4JHH ) 2 Hz, 4JHH ) 2 Hz), 6.16 (2H, dd, 4JHH ) 2 Hz, 4JHH ) 2 Hz), 6.21 (2 H, dd, 3JHH ) 10 Hz, 4JHH ) 2 Hz), 6.20 (2 H, dd, 3JHH ) 10 Hz, 4JHH ) 2 Hz), 6.28 (2 H, dd, 3JHH ) 10 Hz, 4JHH ) 2 Hz), 6.28 (dd, 3JHH ) 10 Hz, 4JHH ) 2 Hz). 13C NMR (C6D6, 125.8 MHz): 138.72 (C5CCH3), 138.58 (C5CCH3), 115.56 (C5SiCH3), 115.36 (C5SiCH3), 114.74 (2-C5H), 114.65 (2-C5H), 110.14 (4-C5H), 110.11 (4-C5H), 108.52 (5-C5H), 108.50 (5-C5H), 32.54 (CCH3), 32.44 (CCH3), 32.38 (CCH3), 0.521 (SiCH3), 0.517 (SiCH3). The variabletemperature 1H NMR spectrum does not exhibit any line-shape changes between 30 and 120 °C, and therefore the population of the meso- and dl-isomer does not change. Anal. Calcd for C24H42Si2Mg: C, 70.1; H, 10.3. Found: C, 70.0; H, 10.6. The EI mass spectrum showed a molecular ion at m/e ) 411 amu. The parent ion isotopic cluster was simulated: (calcd %, observd %) 409 (0, 1), 410 (100, 100), 411 (50, 51), 412 (31, 33), 413 (10, 10), 414 (3, 3). [1,2,4-(Me3Si)3C5H2]Na. Tris(trimethylsilyl)cyclopentadiene (13.18 g, 4.67 mmol) was added to a suspension of sodium amide65 in tetrahydrofuran (70 mL). The mixture was heated to 50 °C and stirred at this temperature for 14 h, ammonia evolved, and the color of the tetrahydrofuran solution changed from yellow to red-brown. The reaction mixture was centrifuged, and the red-brown supernatant was taken to dryness. Prolonged exposure to dynamic vacuum at room temperature (5 h) was necessary to ensure complete removal of coordinating solvents. The off-white solid was washed with pentane (2 × 200 mL) and exposed to dynamic vacuum for 3 h, yielding a colorless powder (12.01 g, 3.95 mmol, 84.1%). 1H NMR (C5D5N, RT, 400 MHz): 7.29 ppm (2H, s, ring-CH), 0.50 ppm (18H, s, Si(CH3)3), 0.31 ppm (9H, s, Si(CH3)3). [1,2,4-(Me3C)3C5H2]Na. [1,2,4-(Me3C)3C5H2]Na was synthesized from 1,2,4-(Me3C)3C5H366 and NaNH2 according to a procedure similar to that used to prepare [1,2,4-(Me3C)3C5H2]Na, yielding a colorless powder in 87% yield. Manganocenes. [(Me3C)C5H4]2Mn. A mixture of [(Me3C)C5H4]2Mg (8.1 g, 30.4 mmol) and MnBr2 (6.54 g, 30.5 mmol) was heated slowly until all the magnesocene was molten at about 100 (77) Schumann, H.; Gottfriedson, J.; Glanz, M.; Dechert, S.; Demtschuk, J. J. Organomet. Chem. 2001, 617–618, 588–600.

Organometallics, Vol. 28, No. 7, 2009 2017 °C. The color of this inhomogeneous mixture changed to red-brown. The mixture was cooled to room temperature, and tetrahydrofuran (20 mL) was added. The red solution was heated to 50 °C overnight; the mixture was cooled to room temperature, and the solvent was removed under reduced pressure. The orange-red residue was extracted with toluene (100 mL). Removal of the solvent and sublimation of the orange-yellow residue at 55-60 °C in a diffusion pump vacuum yielded large red crystals up to several millimeters in diameter, but leaving a reasonable quantity of a yellow-white residue behind. The red crystals were resublimed (40-50 °C, diffusion pump vacuum) to yield 3.8 g (12.8 mmol, 42%). Mp: 59-60 °C (rev); during the melting process and subsequent heating the color of the melt changed from red to yellow-orange. The compound was stable to at least 310 °C without decomposition. Furthermore, the liquid was found to form supercooled liquids, solidifying only after freezing in liquid nitrogen. 1H NMR (C6D6, RT): 12.7 ppm (ν1/2 ) 3100 Hz). The absence of unreacted [(Me3C)C5H4]2Mg was verified by EI mass and 1H NMR spectra. The EI mass spectrum showed a molecular ion at m/e ) 297 amu. The parent ion isotopic cluster was simulated (calcd %, observd %): 297 (100, 100), 298 (21, 20), 209 (2, 3). [(Me3C)C5H4]2Mn has been reported previously by another synthetic route.19 The yellow-white residue from the first sublimation was extracted with hot toluene (60 mL), concentrated to ca. 40 mL, and slowly cooled to -25 °C. Light yellow-green crystals were obtained in 18% yield (5.35 mmol, 2.2 g) and were shown to be [(Me3C)C5H4]2Mg(thf)2. 1H NMR (C6D6, RT): δ 6.32-6.26 ppm (m, AA′BB′, H2,2′ and H3.3′, 8H, ring-CH), 3.65 (m, R-CH of THF, 8H), 1.52 (s, CMe3, 18H), 1.27 (m, β-CH of THF, 8H). [(Me3Si)C5H4]2Mn. [(Me3Si)C5H4]2Mn19 was synthesized by reaction of potassium trimethylsilylcyclopentadienide (7.7 g, 0.044 mol) and MnBr2 (4.69 g, 0.022 mol) in boiling tetrahydrofuran (100 mL) overnight. The light brown suspension was taken to dryness, and the residue extracted with pentane (150 mL). Filtration, concentration of the pentane extracts to 25 mL, and slow cooling to -80 °C yielded yellow crystals, melting upon warming to room temperature. The resulting yellow liquid was distilled, for further purification, in a diffusion pump vacuum (bp 92-93 °C) (3.4 g, 9.6 mmol, 44%). 1H NMR (C6D6, RT): 13.0 ppm (ν1/2 ) 524 Hz). [1,3-(Me3C)2C5H3]2Mn. To a mixture of [1,3-(Me3C)2C5H3]2Mg (2.6 g, 6.8 mmol) and MnI2(thf)2 (3.1 g, 6.8 mmol) was added tetrahydrofuran (50 mL). The reaction mixture turned orange-red, and a colorless precipitate was formed. After stirring for 24 h at room temperature the solvent was removed under reduced pressure and the residual orange powder was extracted with pentane (50 mL). The red solution was filtered, concentrated, and slowly cooled to -20 °C. Two crops of orange-red crystals were collected and recrystallized from pentane, yielding 1.9 g (68%) of the pure manganocene. Mp: 145-146 °C (rev). Sublimation: 55-60 °C (diffusion pump vacuum). 1H NMR (C6D6, RT): 14.5 ppm (ν1/2 ) 2700 Hz). IR (Nujol mull; CsI windows; cm-1): 1293s, 1250s, 1230w, 1197s, 1162m, 1083w, 1048w, 1040m, 1020m, 930m, 920m, 910m, 838s, 800s, 737m, 685m, 655m, 640m, 615w, 604w, 507m, 475br w, 405w, 325w, 300w. Anal. Calcd for C26H42Mn: C, 76.25; H, 10.34. Found: C, 75.91; H, 10.43. The EI mass spectrum showed a molecular ion at m/e ) 409 amu. The parent ion isotopic cluster was simulated (calcd %, observd %): 409 (100, 100), 410 (40, 38), 411 (4, 5). [1,3-(Me3Si)2C5H3]2Mn. To a mixture of [1,3-(Me3Si)2C5H3]2Mg (3.9 g, 8.8 mmol) and MnI2(thf)2 (4.0 g, 8.8 mmol) was added tetrahydrofuran (75 mL), and a colorless precipitate formed immediately. The reaction was stirred overnight, and the precipitate was allowed to settle. The pale yellow solution was filtered, and the filtrate was taken to dryness. The yellow residue was extracted with pentane (30 mL) and filtered. The volume was reduced to 15 mL, and the solution was slowly cooled to -80 °C. Light yellow crystals formed over a period of 10 days (2.4 g, 58%). Alternatively,

2018 Organometallics, Vol. 28, No. 7, 2009 the product may be purified by sublimation (50-55 °C, diffusion pump vacuum). Mp: 90-91 °C (rev). 1H NMR (C6D6, RT): 11.8 ppm (ν1/2 ) 980 Hz). IR (Nujol mull; CsI windows; cm-1): 1293s, 1250s, 1230w, 1197s, 1162m, 1083w, 1048w, 1040m, 1020m, 930m, 920m, 910m, 838s, 800s, 737m, 685m, 655m, 640m, 615w, 604w, 507m, 475br.w, 405w, 325w, 300w. Anal. Calcd for C22H42Si4Mn: C, 55.76; H, 8.93. Found: C, 55.85; H, 8.95. The EI mass spectrum showed a molecular ion at m/e ) 473 amu. The parent ion isotopic cluster was simulated (calcd %, observd %): 473 (100, 100), 474 (45, 46), 475 (23, 23), 476 (7, 6), 477 (2, 2). [(3-Me3C)(1-Me3Si)C5H3]2Mn. To a mixture of [(3-Me3C)(1Me3Si)C5H3]2Mg (2.6 g, 6.8 mmol) and MnI2(thf)2 (3.1 g, 6.8 mmol) was added tetrahydrofuran (100 mL). The reaction mixture turned orange, and a colorless precipitate was formed. After stirring for 2 h at room temperature the solvent was removed under reduced pressure and the orange residue extracted with pentane (50 mL). The solution was filtered, concentrated to ca. 30 mL, and slowly cooled to -80 °C. Small and thin orange needles formed after 2 days at -80 °C. The crystals were dissolved in pentane and recrystallized (1.9 g, 68%). The compound sublimed at 50-60 °C in a diffusion pump vacuum. Mp: 106-107 °C (rev). 1H NMR (C6D6, RT): 19.9 ppm (18H, ν1/2 ) 2500 Hz), 12.3 ppm (18H, ν1/2 ) 1750 Hz). IR (Nujol mull; CsI windows; cm-1): 1325m, 1287m, 1250br s, 1200m, 1170br.s., 1075s, 1045s, 1020w, 935m, 915s, 825brs, 750brs, 687m, 674m, 628s, 480m, 450w, 423s, 385brs, 333m, 290m. Anal. Calcd for C24H42Si2Mn: C, 65.26; H, 9.58. Found: C, 64.88; H, 9.69. The EI mass spectrum showed a molecular ion at m/e ) 441 amu. The parent ion isotopic cluster was simulated (calcd %, observd %): 441 (100, 100), 442 (37, 37), 443 (13, 13). Single crystals suitable for a X-ray structure investigation could not be obtained by crystallization or sublimation. The attempted separation of the diastereomeric mixture by slow sublimation in a sealed ampule at 50 °C over periods of weeks yielded layered crystalline material that was not suitable for an X-ray diffraction experiment. [1,2,4-(Me3C)3C5H2]2Mn. MnCl2 (2.00 g, 15.89 mmol) was suspended in tetrahydrofuran (50 mL), and sodium 1,2,4-tri-tertbutylcyclopentadienide (8.14 g, 31.8 mmol) dissolved in tetrahydrofuran (100 mL) was added. After the addition was complete a yellow solution had formed, which was heated under reflux for 14 h. The solvent was removed under dynamic vacuum, and the yellowbrown residue was extracted with pentane (200 mL). The pentane extract was concentrated to 50 mL and slowly cooled to -20 °C. Two crops of light yellow crystals were obtained in 5.51 g overall yield (10.6 mmol, 66%). The compound sublimed at 90-95 °C in a diffusion pump vacuum. Mp: 308-309 °C (rev). 1H NMR (C6D6, RT): 14.65 ppm (ν1/2 ) 2807 Hz). IR (Nujol mull; CsI windows; cm-1): 3115w, 1580w, 1360s, 1265m, 1245s, 1205m, 1150m, 1100br.m, 1025m, 1007m, 962w, 925w, 821sh.s, 812s, 680m, 545w, 448w, 410w, 380w, 280w. Anal. Calcd for C34H58Mn: C, 78.27; H, 11.20. Found: C, 78.06; H, 11.18. The EI mass spectrum showed a molecular ion at m/e ) 521 amu. The parent ion isotopic cluster was simulated (calcd %, observd %): 521 (100, 100), 522 (28, 28), 523 (5, 5). The synthesis of this compound has been reported recently,50 but some of the stated properties differ from those obtained in this work. [1,2,4-(Me3C)3C5H2]2Mn did not react with bipy, Li(neopentyl), H2 (up to 16 atm), or CO (up to 13 atm). [1,2,4-(Me3Si)3C5H2]2Mn. To a mixture of sodium tris(trimethylsilyl)cyclopentadienide (5.93 g, 19.5 mmol) and MnI2(thf)2 (4.42 g, 9.75 mmol) was added tetrahydrofuran (150 mL). A colorless precipitate formed immediately. The solution was heated to reflux for 2 h, then cooled to room temperature. The solvent was removed under dynamic vacuum and the residue extracted with pentane (150 mL). The pentane extract was concentrated to 20 mL and slowly cooled to -80 °C, yielding ivory-colored crystals (3.05 g, 4.94 mmol, 51%). Further purification was performed by sublimation

Walter et al. in a diffusion pump vacuum (70-80 °C). Mp: 286-288 °C (rev). 1 H NMR (C6D6, RT): two very broad signals, one of which overlapped with C6D6: ∼10 ppm (ν1/2 ≈ 1050 Hz) and ∼7 ppm (ν1/2 ≈ 750 Hz). IR (Nujol mull; CsI windows; cm-1): 3060vw, 1335vw, 1248s, 1135w, 1088s, 1002m, 940m, 835vs, 752s, 688w, 738m, 728m, 504w, 430w, 395w, 378w, 340brw. Anal. Calcd for C28H56Si6Mn: C, 54.46; H, 9.40. Found: C, 54.17; H, 9.43. The EI mass spectrum showed a molecular ion at m/e ) 617 amu. The parent ion isotopic cluster was simulated (calcd %, observd %): 617 (100, 100), 618 (62, 61), 619 (39, 40), 620 (15, 15), 621 (1, 2). Significant amounts of decomposition were observed on reaction with H2 (D2) (10 atm) or CO (16 atm). No deuterium incorporation into the cyclopentadienyl ring was detected by 2H NMR spectroscopy. The compound decomposed completely under an atmosphere of H2 (D2) over a period of 4 days to (Me3Si)3C5H3 and an insoluble brown residue. Attempted Synthesis of {[1,3-(Me3C)2C5H3][1,3-(Me3Si)2C5H3]Mn}. The missing link within the series of tetrasubstituted manganocenes was the heteroleptic representative {[1,3-(Me3C)2C5H3][1,3-(Me3Si)2C5H3]Mn}. In contrast to [1,3-(Me3C)(Me3Si)C5H3]2Mn, this isomer does not form a mixture of diasteromers. However, the synthesis of such a molecule was not straightforward: A solution synthesis was hampered by the fact that high-spin manganocenes with labile cyclopentadienyl ligands were prone to ligand redistribution reactions, which were furthermore driven by the formation of the enthalpy-favored, low-spin [1,3-(Me3C)2C5H3]2Mn species. However, in some cases metathesis reactions in molten solutions were employed successfully to synthesis labile organometallic species, e.g., (C5H5)Mg(CH2CMe3).78 Accordingly, a 1:1 mixture of [1,3-(Me3C)2C5H3]2Mn and [1,3-(Me3Si)2C5H3]2Mn formed an eutectic melt with a melting range 80-83 °C. Heating the sample to 150 °C over a period of 1 month set up a mixture of hetero- and homoleptic compounds as determined by electronimpact mass spectrometry (EI-MS) (mp 74-77 °C). Determining the quantitative composition of the mixture was nearly impossible considering the significant line broadening in the 1H NMR spectra, which precluded an accurate integration. Alternatively UV-vis spectra were hard to deconvolute into the three contributors without pure {[1,3-(Me3C)2C5H3][1,3-(Me3Si)2C5H3]Mn} available. The qualitative composition did not change upon heating for a additional month to about 200 °C (mp 68-70 °C), as determined by EI-MS. The reaction mixture was sublimed in order to achieve some separation (40-50 °C in diffusion pump vacuum). At this point it was of interest to get qualitative insight into the magnetism of the mixture (see Supporting Information for details). As shown in Figure S27, the magnetic moment was clearly different from the one obtained for the homoleptic species. Interestingly, the magnetism resembled the behavior of [1,3-(Me3C)(Me3Si)C5H3]2Mn: the high-spin population stayed constant to ∼120 K, then gradually increasedsmore or less linearlyswith T up to 300 K, but without reaching a saturation value. Unfortunately, the relative proportions of the different species were not available, and therefore it was impossible to determine the contribution of “pure” {[1,3(Me3C)2C5H3][1,3-(Me3Si)2C5H3]Mn} in this mixture. However, the close qualitative relation to [1,3-(Me3C)(Me3Si)C5H3]2Mn suggested that both molecules were electronically rather similar. At low temperature the ground state was a mixture of a constant population of high- and low-spin molecules, but from 120 K the low-spin molecules gradually converted into high-spin molecules. This proceeded in the mixture more gradually than in [1,3(Me3C)(Me3Si)C5H3]2Mn.

Acknowledgment. This work was supported by the Director, Office of Science, Office of Basic Energy Sciences (78) Andersen, R. A.; Blom, R.; Haaland, A.; Schilling, B. E. R.; Volden, H. V. Acta Chem. Scand. A 1985, 39, 563–569.


(OBES), of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231. X-ray absorption data were collected at the Stanford Synchrotron Radiation Laboratory, a national user facility operated by Stanford University on behalf of the DOE/OBES. We thank Dr. Fred Hollander (at CHEXRAY, the U.C. Berkeley X-ray diffraction facility) for assistance with the crystallography, the German Academic Exchange Service (DAAD) for a fellowship (M.D.W.), and Wayne W. Lukens for discussions and assistance with the EPR studies. Supporting Information Available: Crystallographic data, labeling diagrams, tables giving atomic positions, anisotropic thermal parameters, bond distances, bond angles, torsion angels, and leastsquares planes for each structure, solid-state magnetism data of (C5Me5)2Mn, [(Me3E)nC5H5-n)]2Mn (E ) C, Si; n ) 1, 2, 3), packing diagrams for [(Me3C)2C5H3)]2Mn and [(Me3Si)2C5H3)]2Mn, HS f LS

Organometallics, Vol. 28, No. 7, 2009 2019 relaxation for [1,3-(Me3C)(Me3Si)C5H3]2Mn after thermal trapping, EPR spectra of [(Me3Si)2C5H3]2Mn, [(Me3C)2C5H3]2Mn and [(1,3Me3C)(Me3Si)C5H3]2Mn, normalized mole fraction vs T plot for high-spin, low-spin equilibrium of [(Me3C)C5H4]2Mn and [1,3(Me3C)(Me3Si)C5H3]2Mn, variable-temperature Mn K edge XANES and EXAFS studies for [(Me3C)2C5H3]2Mn, temperature dependence of the UV-vis absorption at 269 nm for [1,3-(Me3Si)2C5H3]2Mn and at 430 nm for [(1,3-Me3C)(Me3Si)C5H3]2Mn. This material is available free of charge via the Internet at http://pubs.acs.org. Structure factor tables are available from the authors. Crystallographic data were also deposited with Cambridge Crystallographic Data Centre. Copies of the data (CCDC 700163-700167) can be obtained free of charge via http://www.ccdc.cam.ac.uk/data_request/cif by emailing [email protected], or by contacting The Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB 1EZ, UK; fax +44 1223 336033. OM800922J

Spin Equilibria in Monomeric Manganocenes ... - ACS Publications

Recommend Documents